1. Field of Art
The present disclosure generally relates to the field of digital video, and more specifically, to determining temporal and/or spatial features associated with objects of a digital video.
2. Background of the Disclosure
A typical digital video depicts multiple semantic objects or actions, such as animals, vehicles, stationary objects, and the like. In order to represent one of these objects or actions in a useful way, so that it can be distinguished from other objects for purposes such as labeling the object based on its meaning, features are derived from the object or action based on its unique properties, such as its visual and/or audio data. Selection of the right types of features for representing the objects or actions permits the objects or actions to be more meaningfully represented and thus better distinguished from other objects.